Development of a Tumor Diagnosis Application for Medical Practitioners using Transfer Learning

A brain tumor is the growth of abnormal cells in tissues brain. It affects a large number people different ages, worldwide. Magnetic Resonance Imaging (MRI) most operative and widely used technique for detection because it provides better contrast images However, complexity problem, manual classification process, requirement skilled medical practitioners, huge amount MRI scan data are major factors thwarting timely vs. non-tumor. Early tumors possible by accurately applying machine learning with aim to save time, cost, human life. Recently, deep via transfer techniques was found be highly effective tasks. diagnosis application presented VGG-19-based model knowledge. Five-fold cross-validation demonstrated 88% accuracy along 0.881 F1 score. The could utilized as successful tool aid oncologists radiologists clinical diagnostics process.